llama-server-one: Single executable for multiple platforms. Similar to llamafile, with latest llama.cpp changes.

[BradHutchings]

I call it llama-server-one. It's a friendly fork of llama.cpp.

https://github.com/BradHutchings/llama-server-one

Compare to llamafile. llama-server-one is better at some things, like keeping up to date with the latest llama.cpp additions. It's not (yet) as good as other things, like GPU support.

I welcome the llama.cpp and llamafile projects incorporating any part of my changes. I'll keep the fork alive so long as I have something interesting to contribute there. Your feedback and support is welcome!

-Brad